Climate-model induced differences in the 21st century global and regional glacier contributions to sea-level rise

Abstract

The large uncertainty in future global glacier volume projections partly results from a substantial range in future climate conditions projected by global climate models. This study addresses the effect of global and regional differences in climate input data on the projected twenty-first century glacier contribution to sea-level rise. Glacier volume changes are calculated with a surface mass balance model combined with volume-area scaling, applied to 89 glaciers in different climatic regions. The mass balance model is based on a simplified energy balance approach, with separated contributions by net solar radiation and the combined other fluxes. Future mass balance is calculated from anomalies in air temperature, precipitation and atmospheric transmissivity, taken from eight global climate models forced with the A1B emission scenario. Regional and global sea-level contributions are obtained by scaling the volume changes at the modelled glaciers to all glaciers larger than 0.1 km2 outside the Greenland and Antarctic ice sheets. This results in a global value of 0.102 ± 0.028 m (multi-model mean and standard deviation) relative sea-level equivalent for the period 2012–2099, corresponding to 18 ± 5 % of the estimated total volume of glaciers. Glaciers in the Antarctic, Alaska, Central Asia and Greenland together account for 65 ± 4 % of the total multi-model mean projected sea-level rise. The projected sea-level contribution is 35 ± 17 % larger when only anomalies in air temperature are taken into account, demonstrating an important compensating effect by increased precipitation and possibly reduced atmospheric transmissivity. The variability in projected precipitation and atmospheric transmissivity changes is especially large in the Arctic regions, making the sea-level contribution for these regions particularly sensitive to the climate model used. Including additional uncertainties in the modelling procedure and the input data, the total uncertainty estimate for the future projections becomes ±0.063 m.